RD-A172 991 ENGINEERING PERFORMNCE EVALURTION OF SALL COMMUNITY lAIRPORT MICRONAVE L.. (U) HSI SERVICES INC NASHINGTON DC0 0 RDRS ET RL. RUG 06 DOT/FRR/PM-S6/24

UNCLRSSIFIED DTFROI-82-V-10521 F/S 17/?lllllllllllllEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEU

IM

S.0

MICRC. EOUINTS HRNAINtBRA FSIADRS16* IIii~ ~ 32 32

DOT/FAA/PM-86/24 l7~7-~-v

Program Engineering & Engineering Performance Evaluation ofMaintenance ServiceWashington, D.C. 20591 Small Community Airport

Microwave Landing System (TI Model)1 at Philadelphia International AirportY,

Runway 17

Auus 496Fia Rpr

Az

thrug F tha Repotioa ehianomto

C-:

ThS docurment is Tavpaial on tepbiThdrug theiationitatchncaIno ain

"I

DISCLAIMER NOTICE

THIS DOCUMENT IS "BEST QUALITYPRACTICABLE. THE COPY FURNISHEDTO DTIC CONTAINED A SIGNIFICANTNUMBER OF PAGES WHICH DO NOTREPRODUCE LEGIBLY.

P:- - 4

L-

.o_'-N , N,

% -. "--'

E~*N -e-

*. " . . . . * * ,r . .. °

Technical Report Documentation Page ,,f'

1. Report No. 2. Governmen, Accession No. 3. Recipient's Catalog No. % .DOT/FAA/PM-86/24 ,1WA/ b ~~/ _________

4. Title and Subtitle 5. Report Date

ENGINEERING PERFORMANCE EVALUATION OF August 1986SMALL COMMUNITY AIRPORT MICROWAVE LANDING 6 Perform.ng Organ.zation Code

SYSTEM (TI-Model)AT PHILADELPHIA INTERNATIONAL .AIRPQRT RUINWAY 17 8. Perform.ng Organioaion Report No.7. Author's)

G.D. Adams and C.E. Richardson 1 W Ut ..9. Performing Organization Name and Address

MSI SERVICES, INC.600 Maryland Avenue, S.W., Suite 695 11 Ce.....tor r... o.Washington, D.C. 20024 DTFA01-82-Y-10521,

13. Type of Report and Perod Covered

12. Sponsoring Agency Name and Address

Federal Aviation Administration Final ReportU.S. Department of Transporation ".-__'."

and aintnanc Serice14. Sponsor-ng Agency CodeEngineering and Maintenance Service A - 410Washington, D.C. 20591 APM- 410 .___.15. Supplementary Notes

16. bstruct4% %1 The Microwave Landing System (MLS) program is designed to meet

both civil and military operational needs. It will eventually replace -the Instrument Landing System (ILS). After a system/equipment "4development phase, the Federal Aviation Administration, in 1979 begana Service Test and Evaluation Program (STEP) to obtain experience fordeveloping criteria for siting, installation and preliminary operationa %

procedures. }-Feasibility demo'ground equipment from the developmentphase was used and user aircraft were equipped with MLS receivers.

This report contains engineering evaluation of the FAA TechnicalCenter flight test data taken on the Texas Instruments' manufacturedSmall Community MLS, as installed for STEP at PhiladelphiaInternational Airport Runway 17.

The equipment performance was found to generally meet the FAA-STD-022b Path Following Error and Control Motion Noise requirement, butnot the linearity requirement. Apparent aircraft propeller inducednoise effects were identified.

17. Key Words 18. Dstribution Statement

Microwave Landing System This document is available to the'S i T a El ipublic through the National IService TsTechnical Information Service, I

Springfield, Virginia 22161

19. Securinty Classnt. lOfthis. report) 20. Security Classnt. (of tf,.s page) 21 No f aes 2.Irc

Unclassified Unclassified 62 .

Form DOT F 1700.7 (8-72) Reproduction of completed page authorized

.- 4

-,n.

" "% . '% :" "- , "."."q' ," ''o ', " " ". '" '".rs S C'-" " " "' " .4 "4 " " - " -" . "-- -." - '' . " . """"'

.'...=

TABLE OF CONTENTS

SECTION/PARAGRAPH PAGE

I. INTRODUCTION 1

A. BACKGROUND I

B. GENERAL I

C. GENERAL SYSTEM DESCRIPTION 2 -

II. TEXAS INSTRUMENTS' SMALL COMMUNITY MLS 3• 1>

A. SPECIFICATIONS 5

B. TEXAS INSTRUMENTS' MLS PHILADELPHIA INTERNATIONAL 14AIRPORT TESTS

1. MLS LOCATIONS 14

2. FLIGHT TESTS 17

3. MEASUREMENT STANDARD 17

4. TEST AIRCRAFT 17

C. RESULTS 19

1. CENTERLINE APPROACH 19

a. Azimuth Path Following 19

b. Azimuth Control Motion Noise 24

c. Elevation Path Following C-L 31

d. Control Motion Noise C-L 31

2. RADIAL TESTS 31

a. Path Following Error (Radial) 31

b. Control Motion Noise (Radial) 43

3. Orbital Runs 43

a. Azimuth Path Following 43Orbital

b. Azimuth Control Motion Noise Orbital 47 VON

c. Elevation Path Following Error 47

A'. "N"ij.

•.-- --- .--...

TABLE OF CONTENTS

SECTION/PARAGRAPH PAGEd. Elevation Control Motion Noise

47Orbita l

4. COMPARISON WITH PREVIOUS DATA 47

D. CONCLUSION 54

Ill. REFERENCES 58

.4.:

" .- P.,

,-. -L

-. - ~ ~ - ~ -* * -°

"A a lat.

,.. -..-.. '

!.

FIGURES .Figure Page

1 Azimuth Guidance Set 4 :av:2 Typical Elevation Pattern of Azimuth Guidance Set 6

3 Azimuth Antenna Patterns of Azimuth Guidance Set 74 Elevation Guidance Set 8

5 Typical Azimuth Pattern of Elevation Antenna 9-N

6 Filter Configurations and Corner Frequencies 12

7 ILS and MLS Elevation Location 15

8 ILS Localizer and MLS Azimuth and DME Location 169 AZ; MLS and Tracker; Run 3 Mar 41, 1981 20

CL 3 Deg, App

10 AZ DIFF; Run 3 Mar 31, 1981 21 ,,CL 3 Deg App

11 AZ; MLS and Tracker; Run 4 Mar 31, 1981 CL 4.5 20Deg. App

12 AZ DIFF; Run 4 Mar 31, 1981 23CL 4.5 Deg ApV. (with orbital data superinposed)

13 AZ DIFF; Run 3 Apr 20, 1981 25CL 4.5 Deg App; PFE

14 AZ DIFF; Run 3 Apr 20, 1981 26CL 4.5 Deg App; CMN

15 AZ DIFF; Run 4, Mar 31, 1981 27

16 AZ DIFF; Run 4 Mar 31, 1981 28CL 4.5 Deg App ; PFE

17 AZ DIFF; Run 5 Mar 31, 1981 29CL 4.5 Deg App; CMN

18 AZ DIFF; Run 5 Mar 31, 1981 30CL 4.5 Deg App; PFE

19 AZ DIFF; Run 3, May 13, 1981 32CL 6 Deg App; CMN

20 EL DIFF; Run 4, Mar 31, 1981 33CL 4.5 Deg App; PFE 1. ,

21 EL DIFF; Run 1, Apr 20, 1981 34

___ __"I.:* . -

V

%'g

Figure Page

22 EL DIFF; Run 1, Apr 20, 1981 35CL 3 Deg App -'-

23 EL DIFF, Run 3, May 13, 1981 36CL 6 Deg App

24 AZ; MLS and Tracker; Run 11, Mar 31, 1981 37IOR Radial 3000 ft

25 AZ DIFF; Run 11, Mar 31, 1981 38IOR Radial 3000 ft

26 AZ DIFF; Run 11, Mar 31, 1981 39IOR Radial 3000 ft

27 AZ DIFF; Run 10, Mar 31, 1981 405R Radial 3000 ft

28 AZ DIFF] Run 2, Apr 15, 1981 415L Radial 3000 ft

29 AZ DIFF; Run 11, Mar 31, 1981 42IOR Radial 3000 ft; PFE

30 AZ DIFF; Run 5, Apr 15, 1981 44IOR Radial 3000 ft; CMN

31 AZ DIFF; Run 6, May 13, 1981 456nm CCW ORB 3000 ft (with factory antenna test .data superimposed)

32 AZ DIFF; Run 6, May 13, 1981 466nm CCW ORB 3000 ft; PFE

33 AZ DIFF; Run 5, May 13, 1981 486 nm CCW ORB 3000 ft; CMN

34 AZ DIFF; Run 5, May 13, 1981 496nm CW ORB 3000 ft; PFE

35 EL DIFF; Run 10, Apr 15, 1981 506NM CW ORB 8000 ft

36 EL DIFF; Run 9, Apr 15, 1981 51 -'*'-;6 nm CCW ORB 4500 ft; PFE

V j_37 EL DIFF; Run 8, Apr 15, 1981 52

6nm CW ORB 4500 ft

38 EL DIFF; Run 8, Apr 15, 1981 536nm CW ORB 4500 ft; CMN

• . 5..

Figure Page

39 AZ; MLS and Tracker; Run 7, Aug 21, 1978 559R Radial (at Atlantic City, NJ)

40 AZ DIFF; Run 7, Aug 21, 1978 569R Radial (at Atlantic City, NJ) PFE

41 AZ DIFF; Run 7, Aug 21, 1978 579R Radial (at Atlantic City, NJ) CMN

TABLESPage

Table 1 TI SCAMLS Summary Parameter 10

Table 2 Flight Test Data Runs For Record 18-.

APPEND IX A: DIGITAL COMPUTER PROCESSING/FILTERING FOR ERRORCOMPONENTS

* - b & A - ~ - ' - - .. "

I• INTRODUCTION

A. PURPOSE

The purpose of this program was to test the Texas ..Instruments (TI) model of a time reference scanning beam (TRSB)known as the "Small Community Airport Microwave Landing System"(SCAMLS), in the operational environment of Runway 17,Philadelphia International Airport. The system was previouslytested at the FAA facility in Atlantic City, N.J. (See reference

B. BACKGROUND ","•

Microwave Landing System Development Program. In accordance .Z.

with the "National Plan for the Development of the MicrowaveLanding System," published in July 1971, the United States (U.S.)MLS program is a joint, interservice Department of Transportation(DOT)/Department of Defense (DOD)/National Aeronautics and SpaceAdministration (NASA) development activity, with DOT FederalAviation Administration (FAA) designated as the lead agency. TheNational Plan initiated a three-phase, multiyear developmentprogram to identify and demonstrate a new approach and landingsystem which is intended to eventually replace the InstrumentLanding System (ILS), and is designed to meet both civil andmilitary operational needs as stated by Special Committee (SC)-117 of the Radio Technical Commission for Aeronautics (RTCA) inDecember 1970.

Phase I of the program involved technique analysis andcontract definition. During this phase it appeared that both theTime Reference Scanning Beam (TRSB) and Doppler techniques hadthe potential for meeting the full range of operationalrequirements.

Phase 11, the feasibility demonstration phase, involveddesign, fabrication, and demonstration of both the Doppler andscanning beam techniques using systems installed at the FAA'sNational Aviation Facilities Experimental Center (NAFEC) andNASA's Wallops Station test facilities. The test results fromPhase II were thoroughly analyzed in December 1974 by an inter-service government committee, with full-time participation ofinternational MLS experts from Australia, France, and the UnitedKingdom and part-time participation from other countries. Thiscommittee selected the TRSB technique over the Doppler techniquefor further development and, as a result, the TRSB was submittedto the International Civil Aviation Organization (ICAO) as acandidate for (and subsequently adopted as) the internationalstandard.

Phase III of the program was concerned with the fabricationof prototype TRSB equipment in the different configurationsnecessary to show compliance with the requirements of all majoruser groups. One of these configurations was the TI SCAMLSintended for short runway operations typical of general aviation

A., .

A. ,

-, - J

m..1,,

% -.- -.... 77.......-.-.-- .

- , ,

IZ

_-

requirements and is the subject of this report. "1

Service Test and Evaluation Program. In 1979, the FAA begana Service lest and Evaluation Program (STEP) to obtain theexperience necessary for developing criteria for siting,installation and preliminary operational procedures. Ground MLSinstallations (of upgraded prototype equipment) were completed andoperations conducted at: 1) Washington National (Runway 18 andRunway 33); 2) Philadelphia (Runway 17); and 3) Clarksburg, West -Virginia (Runway 21). User participants in STEP include RansomAirlines and Wright/Aeromech Airlines. Two helicopter operators(Sun Oil and Keystone) agreed to join the program atPhiladelphia. Initially, operations are in VFR conditions only.The operational procedures and criteria are being developed underthis program for use when the first production MLS installationsare made in 1986.

C. GENERAL SYSTEM DESCRIPTION

All configurations of the Phase III TRSB MLS (which is anair-derived system) operate at C-band (5032.0 - 5090.7 megahertz(MHz)). The airborne receiver/processor measures a verticalangle from the elevation transmitting antenna, assumed relativeto the horizontal plane tangent to the runway surface near theglidepath intercept point (GPIP), and measures a horizontal angle ,relative to the runway centerline from the azimuth transmittingantenna. In the TRSB technique, the airborne angle informationis derived by precisely timing the passage of narrow fan beamswhich are scanned sequentially TO-FRO at high rates throughazimuth and the elevation coverage volumes. The time intervalbetween nasn of the TO and FRO beams is directly proportionalto the azimuth and elevation of the receiver and, therefore, theapproach aircraft. Both the azimuth antenna and elevationantenna have a transmitter power output of 20 watts andrespective gains of 14.5 and 16.5 decibels (dBi) relative to anisotropic source, thus providing usable guidance signals out to arange of 15 nautical miles (nmi), assuming a receiver sensitivityof -100 dBm.

Azimuth antenna beamwidth is the major factor in tailoring asystem to a particular runway length in order to prevent inbeammultipath. [In beam multipath is the result of the scanning beamilluminating a reflecting object at the same time it isilluminating the aircraft. The signal arriving at the aircraftvia the rtdlection path can cause noise and errors to beprocessed through the receiver.] A narrower beamwidth reducesthe area where potential reflecting surfaces would be "in-beam".[Current obstruction criteria require hangers, etc, which mi9hthave large vertical reflecting surfaces oriented to reflect intothe approach, to be at least 850 feet from an instrument runway.]The receiver processor can discriminate against most main beamreflections if the reflector is about 2 beamwidths or moreremoved from the beam pointing angle to the aircraft.

Azimuth antenna beamwidth is alsc a consideraticn for the

2

%-%

-- -- o9 . o

oO9,-

vangular accuracy required. Accuracies are linearly specified (infeet) which translates to a smaller angle as the distance tothreshold increases. The basic receiver accuracy performance isa function of the beamwidth. Thus the accuracy requirementsdictate a narrower beam for a long runway. The TI SCAMLS azimuthantenna has a 30 beamwidth, which matches to about a 6000 footthreshold distance. The configuration at Philadelphia had 6567feet between the azimuth and threshold.

One of the design considerations operative in the MLS is theconcept of modularity, in which the system can be configured orupgraded to suit the changing needs of a particular user byadding other subsystems such as flare, missed approach, or rangeas needed at a later time. In addition, most of the electronicsused in the TI SCAMLS azimuth and elevation units can beinterchanged, with some system monitor parameters changed.

II. TEXAS INSTRUMENTS' SMALL COMMUNITY MLS

The TI SCAMLS is a prototype of the system intended toprovide approach and landing guidance in a low cost package torelatively short runways, typical of low-density feeder andgeneral aviation airports, while retaining compatibility withmore expanded versions of TRSB and allowing for growth potential.The system error budget and monitor are designed to support atleast Category I Instrument Flight Rules (IFR) operations (200-foot ceiling and 2,400-foot runway visual range) for runwaylengths up to 5,000 feet.

The TI SCAMLS is comprised of two subsystems; an azimuthunit and an elevation unit. Each unit is completely self-contained within its climate-controlled antenna case and does notrequire additional equipment shelters. Figure I shows theazimuth guidance set which consists of the azimuth electronicscabinet and the azimuth antennas.

The azimuth unit uses a bifocal pillbox feeding a flat-plate 'at .y of 32 waveguides with 37 "C"-shaped slots in each waveguide ..spaced so as to form a vertical fan beam (3 degree beamwidth)Vertical coverage is provided from 1 degree to 15 degrees inelevation with a sharp underside cutoff (13 dB/degree). Thisprototype antenna scans a beam from left 12 degrees throughcenterline to right 12 degrees, providing proportional guidancefrom left 10 degrees to right 10 degrees. Built-in sectorclearance antennas provide full fly-left and full fly-rightcoverage from left 40 degrees to left 10 degrees, and right 40degrees to right 10 degrees. The same antennas provide right andleft side lobe suppression (SLS) signals except that output poweris reduced by 6 dB relative to the clearance signals. The backSLS antenna covers the region -90 degrees through 180 degrees to+90 degrees with 3 dB more power output than the left-right SLSsignals.

3,-.

A . °

-. - . -.- ~ , .. . .'.. . . . . . .

.4P

'.Ju

< z

U'-U

z

*U zAL--

Cc.

a n

% . . .. . . . . . . .

A typical elevation pattern of the azimuth antenna is shownin Figure 2, and the azimuth coverage of the various azimuthantennas is shown in Figure 3. The scanning rate of the azimuthbeam is 13.5 hertz (Hz). The identification (ID) antenna has thesame gain and input power as the clearance antennas. Thecoverage is ±40 degrees in azimuth and from I degree to 15degrees in elevation.

The small community system transmits the following data from .'"

the azimuth unit:

Airport identification (Morse code),Azimuth Status (Category I or unusable), ".-Elevation Status (Category I or unusable),Azimuth offset (lateral distance from runway centerline),Elevation offset,Elevation to threshold distance,Airport identification (digital),Runway identification, andMinimum glide slope.

Figure 4 shows the elevation guidance set consisting of thescanning antenna (40.5 Hz rate), the ID sector antenna, and theelectronics cabinet. The scanning antenna is a bifocal pillboxarray consisting of 12 monopoles feeding a sub-reflector whichfeeds a primary reflector. The antenna radiates a beam 2 degreesin width which can scan from I degree to 15 degrees in elevation.The ID antenna transmits a Differential Phase Shift Keying (DPSK)signal which conditions the airborne receiver to receive thescanning beam that follows. Figure 5 shows the azimuth pattern -.

of the elevation antenna. The TI SCAMLS summary parameters are .-

listed in Table 1.

A. SPECIFICATIONS

The TI SCAMLS was subjected to numerous flight and staticengineering tests as required by the Phase III test plan for theU.S. MLS. The object of those tests was to provide data todetermine if the systems were operating within the accuracy andcoverage limits specified by the Phase III TRSB contracts. Forthe small community system, specification FAA-ER-700-04 applies;with accuracy degradation allowances given in specification FAA-ER-700-07.

As the program has matured, these specifications have beensuperceded by FAA-STD-022b, which is in agreement with therecently adopted ICAO Standards and Recommended Practices(SARPS). For the purpose of the STEP program it is appropriateto use the current standards, even though they may be morestringent than those against which the equipment was designed. .-.

Parameters which may not meet the standards will be highlightedfor corrective measures such as improved design or re-evaluatedstandards.

5. - -',._,

,'° - %°.

'.".,-".._- _i .. .-' '.-''..'-

-. ." ''.,'i"]' 'i; Y '-."."-" '""Y " '," ' ' <"-"-"-" " N \ "..-.-' ---.* ;' -*'%.''

1

'a. ~

N

ISV0'.

a?

-a

-aI-.

N

a.

a-. U

.5.

-a

* U~ Nt.~ <

N

*9i I

zz 0< ~-.

N

4 N AU

"S

J

+ '.5

0 N q. '.0 ~ 0 N ~ -.0 ~ 0 N q. -.0 ~ 0 N ~ -.0 ~ 0(gp) AVM 3N0 S~MOd TLVT3~I

"-.5.

~ -a,.. ..

-5.-a.

.5.

a-

a',

a.

a'

IQI

in .

t.-j-I S t... & . .. P . .. I - .- ;

,p./. ,

5..;

ELEVATION SET ELECTRONICSCAS3INET

ELEVATIONID SECTORANTENNA

RF INPUT r'EL'VATI ON

;01%,"

A-- S%

w... % .

%5

# J.i= ,.

5 I.,. o-

-4b

I -%

.% ~4* .

* '.N, 'i

~44

le- le ,

7'q

=66

L C

% -%.44 N-

-.P 7.- r- d777%~

Flight measurements were made to determine the azimuth andelevation angular errors in the system (i.e., the differencebetween the angle received and processed by the airborne receiverand the true angle at the same instant in time). The guidancesignals are subject to propagation distortion and processinginaccuracies introduced in both the ground and airborne . .equipment. These errors fall into two categories, constant biaserrors and cyclical errors of all frequencies. These errorsinteract with the flight control system in a variety of ways.resulting in two general types of guidance errors: Path FollowingError (PFE) and Control Motion Noise (CMN).

The PFE is that portion of the guidance signal error which

could cause aircraft displacement from the desired course or glidepath. These perturbations fall within the loop guidancebandwidth of an aircraft. The path following error is composedof the path following noise and the mean course error in the caseof azimuth functions, or the mean glidepath error, in the case ofelevation functions.

The CMN is that portion of the guidance signal error whichcould affect aircraft attitude and cause control surface, wheeland column motions during coupled flight, but which does notcause aircraft displacement from the desired course or glidepath.It may contribute to control surface and servo wear, anddiminish flight crew confidence by presenting them with a "shakystick".

The PFE is comprised of those frequency components of theguidance signal error at the output of the airborne receiverwhich lie below 0.5 radians per second for azimuth guidance andbelow 1.5 radians per second for elevation guidance information.The control motion noise is comprised of those frequencycomponents of the guidance signal error at the output of theairborne receiver which lie above 0.3 radian per second forazimuth guidance or above 0.5 radian per second for elevationguidance information. The output filter corner frequency of thereceiver used for this measurement is 10 radians per second.

NOTE: The PFE and CMN are evaluated by filtering the outputof the receiver (see Figure 6). The filter characteristicsare based on a wide range of existing aircraft responseproperties, and are considered adequate for foreseeableaircraft designs as well.

The FAA-STD-022b (Table 5) System Error Limits at theApproach Reference Datum are:

Approach Azimuth ±20 ft. (PFE) ±10.5 ft. (CMN)Approach Elevation ±0.1330 (PFE) ±0.0500 (CMN)

.

.°

.. ., . , , %"4"

S. -n! ..

4ILRECEIVE"

RECEIER PAH4-ATFOLLOWING

RA NL RO OLWN

SREFECEGUDNE CONRFEUECE ROIN/E)RECEIVER OTPAT FLE

APOCPAHFOLLOWING PAT2

AZIMT 0.5S 0.3R10 INL~TE S2+2 n S

ABSLUT ___________CONTROL__ MOTION NOISE FITR MOTIONs

FRENCE . 1 Sw

GUIANE CRNA FEOENCES(RAIAS/CIRECEIVER OUTPUT FILTER (LO PASS

J PRAHPATH FOLLOWINGRLOPAS

N ELVATIN 15 0. 10 CONTROL MOTION NOISE FILTER S 1.56PASS)

FLAR 2.0 0.5 0 S0

REEIE OUPU FITE (LOW ASSJ-p..

FigurePAT FOLLOWIN Co fitERn aLOW PASS Feqenie

CONTOL MTIONNOIS FILER (IGH ASS

12

%21kI% L % r %

The approach reference datum is 50 feet above threshld onthe minimum glidepath. From the reference datuqm to the coveragelimit, the PFE and CMN limits, expressed in angular terms, areallowed to linearly increase as follows:

(1) For azimuth functions:

(a) With the distance along the runway centerlineextended, by a factor of 1.2 for the PFE limits and to ±0.10degree for the CMN limits.

(b) With azimuth angle, by a factor of 1.5 at the ±40degrees and a factor of 2.0 at the ±60 degrees azimuthangles for the PFE and CMN limits.

(c) With elevation angle from +9 degrees to +15degrees, by a factor of 2.0 for the PFE limits.

(d) Maximum angular limits. The PFE limits shall notexceed ±0.25 degrees in any coverage region below anelevation angle of +9 degrees nor exceed ±0.50 degrees inany coverage region above that elevation angle. The CMNlimits shall not exceed ±0.10 degrees in any coverage regionwithin ±10 degrees of runway centerline extended nor exceed±0.20 degrees in any other region within coverage.

NOTE: It is desirable that the CMN limits not exceed ±0.10degrees throughout the coverage.

(2) For approach elevation functions:

at) Wit distance along the runway centerline extendedat he iniumglidepath angle, by a factor of 1.2 for the

PFE liisadto ±0.10 degrees for the CMN limits.

(b) With azimuth angle, from runway centerline extendedto the coverage extreme, by a factor of 1.2 for the PFElimits and by a factor of 2.0 for the CMN limits.

(c) With increasing elevation angle from +3 degreesto +15 degrees, by a factor of 2.0 for the PFE limits.

(d) With decreasing elevation angle from +3 degreesV (or 60% of the minimum glidepath angle, whichever is less)

to the coverage extreme, by a factor of 3 for the PFE andCMN limits.

(e) Maximum angular limits. The CMN limits shall notexceed ±0.10 degrees in any coverage region, within ±10degrees laterally of runway centerline extended, which isabove the elevation angle specified in (d) above.

NOTE: It is desirable that the CMN limits not exceed ±0.10degrees throughout the coverage region above the elevation

Nj angle specified in (d) above.,.I,

3' 13

I-o

m," ..................................

is r this equipment, tr'e design coverage of the azimuth unit-in range, ±10 degrees in aziuth angle, and 1 degree to

15 degrees in elevation. The design coverage of the elevationunit is 15 nm in range. ±10 degrees in azimuth (relative to the

azimuth site), and 1.9 degrees to 10.67 degrees in elevation.

The calculated accuracy specification limits for the threetypes of flight patterns flown against the TI SCAMLS are shown onthe data plots. Both the contract hardware specification limits.%and the FAA-STD-022b specification limits are shown. The threetypes of flight profiles are centerline approach, radials, andorbits. The curves are plotted only out to 8 nmi because lasertracking beyond this point was not considered highly accurate,usually due to weather conditions during flights.

NOTE: For FAA-STD-022b purposes, the PFE and CMN forapproach azimuth or for back azimuth shall be evaluated overany 40-second interval of the flight error record takenwithin the coverage limits. The PFE and CMN for approachelevation shall be evaluated over any 10-second interval ofthe flight error record taken within the coverage limits.The requirement is interpreted to be met if the PFE or CMNdoes not exceed the specified error limits for more than 5percent of the evaluation interval.

B. TEXAS INSTRUMENTS MLS PHILADELPHIA INTERNATIONAL AIRPORT

1. LOCATION. The Texas Instrument Small CommunityAirport MLS (SCAML was located to serve Runway 17 atPhiladelphia International Airport. The Morse code identifierwas XZY. The assigned MLS channel was 596 (5059.8MHz). An ILShas also been installed to serve Runway 17. Although Runway 17has a 743' displaced threshold due to obstacles outside ofthreshold, the elevation site is located to give a 55' crossingheight over the normal threshold (beginning of the pavement).This may place the obstacles closer to the elevation beam lo w erlimit than would be normal. Figure 7 shows the MLS elevationlocation 250 feet from centerline and also the ILS glide slopeantenna location 538 feet from centerline. .

The azimuth site is located on the centerline extended(Runway 17/35) 1100' beyond the stop end of Runway 17. This putsit well below the clearance surface for Runway 35. The nearestbuilding to the runway is at an azimuth angle of about 7 degreesfrom runway centerline and at a maximum elevation of about 0.7degree. In this position it should not cause noticeable .interference on the approach path for either the elevation orazimuth signals. Figure 8 shows the MLS azimuth location oncenterline with the DME antenna located 400 feet off centerline,and the ILS localizer located 245 feet behind the azimuthantenna. 7o

14

building t the runwa..s.at.an.zimuth.an.e.of.abou.7.degrees-........frmrna etrieada axmmeeaino bu .

Ft .

LIS 34LpL

Elvaio

FIGURepo~ E 7ISadMSEeainLct

153 Ft

.71

MLS DME1

35 Aimt

* 4)

40 Ft..(N

/L oaie

IA-

FIU E 8 IS L c l z r M S Az mt n M o., o

As part of the normal ILS installation procedures, anFAA flight inspection crew and aircraft measured the performanceof the ILS against the ILS flight inspection commissioningcriteria. The MLS elevation and the ILS glide slope antennas areabeam one another but separated by 288 feet so no interferringeffects should be anticipated. The azimuth antenna is 246 feetdirectly in front of the ILS localizer. The flight inspection of9/20/83 reported the ILS facility operation was satisfactory andgave it an unrestricted rating. [Note: At some otherfacilities, the ILS has been inspected before and again after theMLS was located in front of the ILS. It has been found thatobjects placed in front of the localizer symmetrically about therunway centerline do not have a serious affect on the localizerperformance. In this case at Philadelphia the MLS was installedfirst, so no "before" data are available o. the ILS.] The ILSwas located so it would have no affect on the MLS performance.

2. FLIGHT TEST. The tests were divided into threeflight pattern types: centerline approaches; constant altituderadial flights; and constant altitude orbital runs. Thecenterline approaches were made on 3-degree, 4.5-degree and6-degree glidepaths to test the elevation and azimuth guidancealong centerline.

The radial runs were designed to keep a constant 3,000foot altitude and a constant radial direction from the azimuthsite. Flights were flown inbound on centerline and ±15 degreeradials from the coverage limits.

The orbital runs, made at altitudes of 2,000, 4,500 and8,000 feet and at a constant 6 nm range fror azimuth,also test the limits of coverage. The azimuth error plotsnormally can be used as a good indication of azimuth pointingerrors at various elevation angles. The flight test runs arelisted in Table 2.

3. MEASUREMENT STANDARD. The true spatial position ofthe test aircraft was considered to be the position measured bythe FAA's laser radar tracking system. The system, built bySylvania, uses a pulsed laser beam to track a retro-reflectormounted on the aircraft. The horizontal angular, vertical '- '

angular and distance coordinates measured from the trackingsystem are translated by the system computer in to a coordinatesystem relative to the MLS antenna phase centers. A correctionis made for the displacement of the retro-reflector from the MLSreceiver antenna on the aircraft.

4. TEST AIRCRAFT. The flight test aircraft was anFAA twin engine, turbo prop Convair 580 based at the FAATechnical Center, Atlantic City, N.J. The MLS receiving antenna

* was an omni-directional stub mounted on the aircraft nose infront of the windscreen. The laser retro-reflector was mounted

17 ."., %

"...............................................

%!

Flight Test Data Runs For Record

TABLE 2

,-A

CENTERLINE APPROACHES

Date Run Glidescope

31 March 1981 1 302 30*

.9 T

" 4 4.50,, 5 4.50

04 April 1981 1 30" 2 30- 3 4.50" 4 4. 50

5 606 60

20 April 1981 1 30" 3 4.50

13 May 1981 3 60

RADIAL APPROACHES

Date Run Radial Altitude

. .

31 March 1981 6 1OL 3Kft" 7 10Li 8 5L

9 010 5R11 1OR

15 April 1981 1 10L 3Kft*2 5L

" 3 04 5R5 1OR ".

20 April 1981 7 10L

PARTIAL ORBITS

Date Run Direction Ranae Altitude

31 March i981 12 CW 6nm 2KftV5 " 13 CCW U ""

15 April 1981 6 CW " *I

7 CCW I

8 Cw 4. 5Kft" 9 CCW

10 CW 8Kft11 CCW

*Indicates bad run 21

,..

. ..H'

on the aircraft top over the cockpit area. The MLS installation Perconsisted of two receivers mounted in racks in the cabin. About30 ft. of one-half inch diameter semi-rigid coax and RG-214brought the signal from the receiving antenna to the receivers.A signal splitter was used to feed both receivers. Thisinstallation results in about 6 dB loss between the antenna andthe receiver. The digital output of the MLS receiver wasdigitally recorded on tape along with time of day and otherparameters of interest for later correlation with data from thelaser tracking system.

C. RESULTS

1. CENTERLINE APPROACHES. Twelve runs were evaluatedwith glidepaths ot 3', 4 1/2 and bu. All exhibited lasertracker loss for 1 to 1.5 nm between 6 and 8 nm from the azimuthsite; however, no obstacles were apparent on the ObstacleClearance Chart to account for this loss, especially at higherglidepath angles. In most cases the data were valid from thereto threshold, although on some runs there was another lasertracker loss at about 4 nm.

a. Azimuth Path Following (C-L). There appearsto be a correlation between the low Trequency component of theazimuth error and the approach course deviations. Where coursedeviations are great, the low frequency component is great. Runs3 and 4 of 31 March 1981, are examples of this characteristicRun 3 is a fairly straight flight with deviations from about+0.45 degree to +0.1 degree during the major portion of theapproach as indicated in Figure 9. The azimuth error plot showsa very low amplitude low frequency component (Figure 10). Run 4..--deviates (over the same range) from +0.7 to -0.5 degree (Figure11) causing a very large low frequency component in the azimuth

error plot (Figure 12).

The initial assessment was that the correlationwas related to the tracker or the filtering applied to thetracker data. However, it appears that the major factor is theantenna beam pointing error. See Figure 31 for a composite plotof an orbital flight at 6 nm, 3000 foot altitude with the factoryantenna test range data at a 30 elevation angle superimposed.There is good agreement. Orbital error data have been extractedfrom the Figure 31 orbital flight data and superimposed on theapproach error data of Run 4 of 31 March 1981. See Figure 12.[The deviation angle from centerline was obtained from Figure 11which shows the aircraft position about centerline. The anglewas used to enter Figure 31 to read the orbital error data.]Figure 12 shows the orbital and approach error data in closeagreement.

The azimuth path following error was within bothoriginal hardware and FAA-STD-022b tolerances for all centerlineapproaches even with the antenna beam pointing errors.

19

.'. A -J

F . . , , . , . . , .. ., ,. , , , , . . . .- .. , - ,, , ., .. . . . . ,. , ., , . ." ,. '. . ' ' . -, ., -. -, , - . ., -, , , , ,. ', .' . ", - - . , .. ., '. '. ,- . - . , , ., -, ,

Li C,

-

C,,

C-0

;cr

0-0

0..20

- -

C2

CL. U- C?

cc

Li U

C; C;

D 21

1.id3S 93~

uj -a-Ji

Li'S

c11.

U- C22 0~V..

-P-

r c

El Ix

'~C x. ,

ri CCC

i V3 "17 ~w e2 2. .7 ; .

L'K Ob

1ES1 d3S 9* r -j

cc I

LI

-*c.. aCL~

C-

4aa-a.

aca

or CC *,

T. T.L

.0 c .0 It 90~ C - .0

It) 0 n c (

I?. C; C C C- - l 0J-C.0.C

23~

'Z' C'

These low frequency components affect the path • "

following error and control motion noise plots. The PFE filterwill respond only to the lower frequency components of the rawerror. However, some of the lower frequency effects are alsoreflected in the control motion noise plots.

._ .% ,.

Figure 13 is a plot of the azimuth PFE for Run 3 of 20 April1981, showing the lower frequency effects. Figure 14 is the .azimuth CMN for the same run. It can be seen that the lowerfrequency components cause the CMN to exceed the tolerance limits,and although actual peak-to-peak levels of the higher frequency"noise" are large but they would be within tolerance if thevarying bias component is removed. The large excursion at about5.5 nm must be considered a tracker fault and ignored since thetracker was not locked for the previous 1.4 nm and was locking onthe reflector again at this time.

b. Azimuth Control Motion Noise. The azimuthnoise was high on the lower glideslopes, moderating only slightlyon the higher (6 degrees) glidepath approaches. If the majorcause of the noise was vertical obstacles, the noise would havemoderated more, and as noted previously, the nearest verticalobstacle, at 7 degrees from centerline, would be considered out . .._of beam. The noise was significantly less at greater ranges ascan be seen in Figure 15 which is a 4.5 degree glideslope %

starting at 10 nm from the azimuth site (Run 4 on 31 March 1981) .

Early in the MLS development program, the errorsdue to signal blockage by, or reflection from, propellers wereinvestigated. It was concluded the reflected interference under 4'these conditions would only cause small amplitude variations inthe received scanning beam envelope. More recently, noiseeffects have been noticed in certain MLS flight data collected onwider beamwidth systems installed to support the MLS STEP. Thepeak errors recorded are not consistent with signal-to-noiseratios in the receiver and become particularly noticeable after achange of MLS antennas on the CV-580 test aircraft. The MLSreceiving .,,tenna for these flights was an omni-directional stubmounted,-46 the aircraft nose in front of the windshield. Thisturbp-'prop aircraft has unusually large propeller blades, and thety ical propeller rotation rates are high enough to cause aSignificant relative phase change (and thus a distortion of thedirect signal) during the dwell time of the wider scanning beamenvelopes. The effects appear to be function of the propellerpitch, becoming noticeable when the pitch is set for approachpower settings. This may be the effect shown in Figure 15 withthe lower noise levels at greater ranges corresponding to cruiseconfiguration. This is further supported by noting the lowernoise levels on the level flyovers, (Figure 26, 30) which wouldbe in cruise configuration.

' ' m . m . J--' 4 . .. --. o. ......... , .. - -. - mm., . .m-~ -

•~~~~~~~~C .- _._,_ Z..' ..

~ ~ * 4 ________________________2

* - -- .. .. .. .. .. .. .. .. .. .. .. .... .. .. ..

93?

JM v VAi V P

1.0

ILw 0

I cCC?

-l 0 N C

0c~ CDL - ;r

Cr

%

C,

cl

C?)

,1,c NIJo Oc25

IAf

ce

'U"2

t CC-C

wo

0 W(I, 6)

- -N

-. I

4-. <

IC.,

II I SIM

C; a 0 C aC7C.; c?4? 0 ~

91SI AJ I--i- Z

'26

p.0 crcr -i

"-. -.J 'v

CLC

C..3

-- -

p~ll

C-,

z3

C4. tC, C,

SIA Ioa N'l 1'0

p27

.3cl crcc -J ..... j U.

L- C) .J

44 4

c.- I C e.c. 4= ~0

C.4 E- . 9

CC

-C?

ef .C; .C; ;cl~~~~~- 12 C

-.c l ' .,'

28..

a,* -A

E-r Lo

0 %-

, 31I

.2 ;-.:Z-i ~ -1

a-- a- -

IDiI -. cc

a- Lt.

. . . ..... .....

3 f Uj -1 A- 10Z-

* -J J I29

I , u

9L I

-j-E- V)

cc U')'

IL.

a.l

%- 5.'- W1 C)

'-430

-r - . r, .- -- ' r,-' .',-'.e ' ' .'- p .

.-

-' -

' " " - " - '

Z" " " -' - "- o. .- -. ". "- ° -- - "-"..

The azimuth Control Motion Noise (CMN) data areconsidered not to meet the hardware or the FAA-STD-022bspecification limits, having too many points outside the limits.However, the data are also considered to have a high likelihood ofbeing contaminated by propeller multipath effects. The 60approach data shown in Figure 19 come close to meeting '

requirements.

c. Elevation Path Following (C-L). The pathfollowing error was well within limits on all runs (and glidepath angles) with higher relative amplitude at the lowerelevation angles and low relative amplitude at the higherelevation angles. Some noise appeared in the path followingplots as indicated in Figure 20, Run 4 of 31 March 1981.

d. Control Motion Noise (C-L). The elevationCMN has low frequency components that ten-T -drive peaks(probably exceeding the 5% allowed) out of the establishedtolerances at the lower elevation angles. This can be seen onFigure 21, Run I of 20 April 1981, where the low frequency (bias) -

takes the form of a ramp function during the middle section ofthe run. However, the ramp may due to the CMN filter response tothe one mile break in the data and the amplitude change fromnegative errors to positive errors (Figure 22) during the break.

As the elevation angle increases the controlmotion noise moderates from peak-to-peak of 0.10 at 30 elevationto about 0.050 at 60 elevation (Figures 22 & 23). However, asignificant portion of the CMN is believed due to the propellerpitch during the aircraft's approach configuration causing amultipath reflection forward into the omni receiving antenna.

2. RADIAL TESTS. Thirteen radial runs were evaluatedon centerline, on five degrees, and on ten degrees off centerline(left and right). These runs were all made at 3000' altitudefrom about 10 to 4 nm or about 7 degrees elevation angle from theazimuth site at the minimum range.

The results were similar in nature to thecenterline approaches. With the wider tolerances, all test dataare well within requirements.

a. Path Following Error (Radial). The pathfollowing error was within tolerances for all radial runs. Someof the plotted PFE can be attributed to antenna pointing errorsand to the aircraft flight path corrections. This can be seen inFigures 24 and 25 which represent the tracker vs MLS and azimutherror plots for Run 11 of 31 March 1981. The effects were .-reflected in the CMN as indicated in Figure 26. A positive biaserror of about 0.060 was evident on the 50 right radial while a -

negative error of about the same magnitude was present on the 5cleft radial. Figures 27 and 28 illustrate this bias. Atcenterline and 100 on either side of centerline there was nosignificant bias.

-7-'

* **: .w~:.?&i~'~:-*.* I.

N -J

VI I

4p 0 11

Li I.:- -

J&O I C

-10

C.C.

-j-

o co

C.)

L.-

\ C; C; a

0~~~ 0 3C-0 c

I'a Za

32O

44

'It.

a. .

7. C)I . - (in

LQ . I

CC

7C

4.- C)sl Zlo W Aa

C,'

I 33

LU'. .

II-

41 I

cr, C C..

L. C3 u c

0

CC2

4- . -

TI.I

0~

.I0

0 co

CN,

- C?

IA h

C?)

-434

% -F

d.%

,- -

121

C,

oto

Co cl 1 3 -.

I -

35.

-I. A

%J. .0.

,-dL

e -J

I L.

.~. -- f

0s

*. *-0 cl c o

zN

36i

- - - F- ~ ~ ~ '2

C.- -.LiL

- U~ l

c.r I.

U. C ~ Irill

7C

C- C4

CC

11~~~ cl0 ' cl

4p Ll-~ON6 SIW C,

37~

% % % %

.- A

'a,

'a C?Cr V)

- 4n.

C 3J

a. CF

cc

00

0 0

0 U ~-0 N -

7.J

0 0 0

a4 0 D a C C3 s1 6o -' '(

380

A- -A P: !W - -. p -F -.- x- -- -VI 0 .. m

%p ~

aEl

403 (' c.

Cg

q

C3 '*5

C- 2

--- 4

% %

I 0

C)-)

E-. U) C 4

-f-- C4co

- ,5 0 c-C;

Il I I)

r2' 1' 1'394

% -J

-i J

r ~C '.I--

r -L "

CI--

ftm C= E-4

C3,

0,

fta Zwe

-4

-; C- 0

o3 Z

p4

40

% ir

. •' aj'

zL L 4

.1 "o , J^, .

0 .0.

% %)

or.- I.-

a: ca

oo"CL

c, I--

C7o

o .o

C.7

" 0; -;° ° " 30] N'

.u '-4

a c

C3

% %

C2 -

ao

060

0

'a. ' .. -: .

414

% Id

r-f e-e- r

-~~~~~~~m -t - . c-r. ~ '

c m It. cl

41 V

0~ CCl

< 0

t 11

0 -4

a- e4A

% 00*0c

C. I- t' S T--4"1 ' -1

424

JIa

-* . - ..-- -.- r -*. -r -r- .- . . - -. "

' .

b. Control Motion Noise (Radial). Controlmotion noise was within specs for all radial runs, althoughhigher in amplitude than when previously tested at the FAATechnical Center. Some low frequency error effects are evidentwhich cannot be considered "noise", similar to that evident onthe centerline approaches. Figure 30 for a 10 degree R radial isa typical plot of CMN with a low frequency effect. The noiselevel on radials is about 1/2 the level experienced duringapproaches and is considered due to the different multipathenvironment (propeller path) between level radial and approachdescents.

3. ORBITAL RUNS. There were 12 orbital runs made,all at a range of b nm, at altitudes of 2k, 3k, 4.5k and 8k feetyielding a range in elevation angles of from 3 to 15 degrees fromthe azimuth site. The runs were both clockwise (R-L) and counterclockwise (L-R). These runs test the coverage requirements andindicate beam pointing errors. Two counter clockwise runs out offour runs at 2000' lost track (tracker) at about 60 beforecenterline until well after centerline.

At the higher altitudes, the aircraft was out ofelevation coverage and thus indicated a very large elevationangular error. One azimuth plot at the 8000' level (above 15degrees) indicated a weak signal by a number of frame flags,apparently due to the beam pattern rolloff.

The scale, measured in time, is greatly different onthese plots than on either radial or centerline approach runs.Assuming the same speed (e.g., 250 '/sec) the aircraft movesalmost 4 degrees on the orbital plots while only about 0.4 nm in10 seconds, as scaled on the radial and centerline approachplots. This is a factor of about 4 to I and makes the orbitalnoise frequency appear much lower.

a. Azimuth Path Following (ORB). The orbitalflight patterns measure the azimuth beam point-ing errors. It canbe seen from the azimuth error plot, Run 6 of 13 May 1980 (Figure31) that guidance down the -1 degree radial would yield a 0.175degree PFE. For reference, the pointing errors measured at the .factory at 30 elevation have been superimposed on a flight dataplot, showing good agreement. Radial flights were run oncenterline and 5 and 10 degrees on either side of center-line,where the errors are indicated to be within tolerances.

The PFE were within tolerance for all runs, deviating ........at times due to the data processing filter initialization. Thiscan be seen on Figure 30 which is the PFE plot from Run 6 of 13May 1980. Referring to Figure 31, it can be seen that the errorplot (at 10 degrees) does not indicate such a deviation withinthe coverage volume. Where these deviations occur, the test 7-'aircraft is always first entering the coverage area.

43

4077 -Y -.

pr

I.z

0

L QI.3 CA:f

LLa

C)CCD

II I

E' LO M' V

C,

Q~~~~ 410P4- r

44r~

-i -

z z

a.. E- - L

.- , .

.*4 z-

<0 E-C

4.-5

p * 0Aft

4%

4.'~c

-JW'

CL,

cc ia

CL I.

4.C

4.o

4k1

U 0?

0 C)444

E- U)i

Il 6, r, 4

-Z-

~~.2:A -: .. ; ..- -. ..... ' - -.- -~- _ . . .- -~ -

Around centerline, where the 0.10 error per degree of anglelinearity requirement (FAA-STD-022b) applies, large error changesare noted in Figure 31. These are calculated as 0.180 error perdegree of angle or larger.

b. Azimuth Control Motion (ORB). Figure 33 fromRun 5 of 13 May 1980 indicates two out-ot-tolerance conditions ona single cycle of about eight seconds duration. This would berepresentative of a noise frequency of 0.13Hz, about 0.8radians/s. This is well within the CMN band which includesfrequencies from 0.3 to 10 radians/s. These excursions are dueto azimuth beam pointing errors.

C. Elevation Path Following Error (ORB). ThePFE were all within tolerance except for the 8000' runs wherethe aircraft was out of coverage. At the 4500' level, theelevation had a positive bias but was still well withintolerances as indicated by Figure 36 for Run 9 or 15 April 1981.The "clean" appearance of the error is typical of the orbitalruns.

d. Elevation Control Motion Noise (ORB). Thecontrol motion noise was within tolerance for all runs. Almostall the raw error frequency spectrum was included as CMN as canbe seen from Figures 37 and 38 of the raw error and the CMN forRun 8 of 15 April 1981. The CMN indicated a positive bias inFigure 38, probably caused by early initialization but should notbe considered a part of the "noise".

Elevation noise appeared to be at a lowerfrequency than the radial or centerline approaches. However,taking into account the time scale differences, it was computedto be about the same. The noise moderated somewhat nearcenterline and the bias error was low at the important lowerangles (3-6 degrees). Some of the higher angle orbits had noerror plots, caused by the aircraft being above the highest scan ..-angle. However, apparently the aircraft was receiving and .-recording the highest scan angle. Figure 35, Run 10 of 15 April1981, indicates the types of errors encountered under thesecond iton s.

4. Comparison with Previous Data. Figure 31 (page45) shows orbital flight data for the azimuth. Also plotted onthat figure is the factory antenna ranne ta +ta an "-',elevation angle of 30. This is a degree or two below theelevation angle on the orbit, but the two sets of data show agood correlation. The lapsed time was about 4 years and the . -system was installed at the FAA Technical Center before being

* moved to Philadelphia.

47

....................... .... ................ ........ . . .• .

.... ............. .•,. ...•. . .. •.... .. . ........ ...... .

cc -i

go, .2*

0n %L~LI

< I -,11

C2 4

cr~ b U

a cc

0 l.A

0 C2 0

4.4 8

* - - p . .-.- I.K b-Mr.... -. -0 - .1 %, 7

CC,

*6 (5

(5 O

Cc -w c

1, 0 0 6~**

- a-4

ILII=CI

41. 0 0~

M 0n

(N o

~e- o C

C 0;

C; C;C;);

C.,/U00Nl 'C:49

186L 130 1

0 cr

Lii ad.

U- C3 -

IL I.-C22

0C2u

*cl

C?<

W; r4 td

soo

.. . . . . .

3t 130 Z

Lic I

C))__ _

r~ 1. 4

CL WI II L

- E -4

t*44

2 -.I7

c.I. I L

C-- , 0 C?a.1 C;clC

o~r51

'SEA '30 1crCr

-JOJ4 -J I-W

7j -t %~

cloa-

cc %" -c

211

0

cc I-

M CrC2

C2

.4-44

jo

W

0 %

CS tsC, 0 ; ; % c -

a a**pC2S 71 6j '10 N6 I I'C t'A' ~ 4

52

*~~~ . .d '. . . . . - .. .

U44.

9L

'4o 'J

CLM)

CL P-..4.Q V

C, I2

.4 .-. )-'C.

a.-P %- I

o a.

0 C;C C, C,-

53

d M T

A -- --

NI

The Philadelphia flight data are noisier than the flightdata taken at the Technical Center. The radial data are taken as .an example because of the apparent propeller induced noise on theapproaches. See Figures 24, 25, and 26 (100 R radial) atPhiladelphia and Figures 39, 40, and 41 (90 R radial) at theTechnical Center, N.J. The source of the extra noise is notidentified here. Some factors may be:

the system reliability seems to be low andproblems difficult to find and fix;

different installation/environment; A.

different ground based tracker (theodolites atTech Center, laser at Philadelphia; and

some change in the airborne MLS receivingantenna/installation.

0. CONCLUSION

The Texas Instruments MLS installed at Philadelphiaprobably met the FAA-STD-022b performance requirements for PFE

and CMN, if the apparent propeller induced noise effects arediscounted. However, one requirement not met is for linearity,which (within ±0.5 degree laterally of centerline) requires theslope of the mean angle errors shall not exceed ±0.1 degree errorper degree of angle.

5%

9..;.".",-, . ... .- .,. ,,..,'.... '.- - ./ - -. . _; ','-','-: '- - . ,., ..- ,-,. . . . , , ..

0 p ." ._, _ "_. _. r "_.-, . . . . . . . . . . . . . . .

ca c

-'".,I.,-.

C *

13 ccC

'- ,.- " t--4qr

'5'0'

-- 5 .- ., . -

E--

z -

,.. ++ 5.s

~z

2571 PH 115

. .,.:.-'--:

0-..

• - S .,""

,,<

"".,, ". " ,

,, "d:. ,

L4.

l.-A.-

* ..\ ,. ...,,

L ~~i. ,-

. . .. .'." "b'.",,',.",

:,+J,+ H+ + tL P

~Z~'2 -. ~;~- ~- ~- -'- *- ~ *:-:.|>: Kl.- -y-:;. -____

cC

IL e-

CL)

-4D

.0

Ix-F.4

4-4 z -

C,

C3 0 r7 0 C) CN'l A ,13 z25/TI HI 5z

Cc.

or -

-J is-C?

I- 0A L/

0r-4

z -

-. In

a,,-,

62 f

o) D

C) El 0 C

25/T .PH 57*~

II. References

1. Test and Evaluation of Texas Instruments Small CommunityMicrowave Landing System, Report No. FAA-RD-80-49;, FAA-NA-79-34; May 1980 John Warren, NAF'EC, Atlantic City, N.J. 08405

2. FAA-STD-022b: October 27, 1983; Microwave Landing System(MLS) Interoperability and Performance Requirements

3. TI. 6850.18; Instruction Book; Small Community MicrowaveLanding System (MLS)7 Texas Instruments Incorporated.

.. vA..

Ir

,1%

. . 91 P. V 7- I , .7 7.

.A-- .. .-

-* APPENDIX A

DIGITAL COMPUTER PROCESSTNG/FTLTERTNG FOR ERROR COMPONENTS

The transfer function of the analog low pass filter used to

extract the Path Following Error (PFE) from the raw data is:

H(S) Wn2/(S2+2WnS+Wn 2)

where, for AZ:W n = 0.78 rad/sec and

for EL:Wn = 2.34 rad/sec

Implementation of this analog filter for computer processing isbased on approximating an integral by the trapezoidal rule and Z-transform theory ("Digital Signal Processing," A. Oppenheim andR. Schafer). By making the following substitutions, thedifference equation for the corresponding digital filter willresult:

2 (l - Z-l)

T(1 + Z-1).

Y(Z) = H(Z) X (Z)

Xn-1 = X(Z)Z-i

Yn-l = Y(Z) Z-1

where the Y's are the calculated filter outputs and the X's arethe measured input values.

T is the sampling period (assumed constant)

Yn = (4+4WnT+Wn2T2)- (Wn2T2) I(Xn+2Xn-l+Xn-2) +

(8-2Wn2T2 )Ynl-(4-4WnT+Wn2T2 )Yn-2"

AZ: T = 2/13.5EL: T = 2/40.5

The filter is started by initializing all values to the firstangular error difference measurement.

After the data filtered, they are compared to the 2-sigma maximumspecification limits.

A-i

,. . . . . . . . . . .. ,

,-

These Control Motion Noise (CMN) errors are generally of afrequency too high for the aircraft to track, but low enough forthe control system to respond to. Thus, CMN results in rapidsmall-amplitude control surface shell and column motions and isundersirable in that it contributes to control surface and servowear and diminishes flight crew confidence by presenting themwith a "shaky stick". The transfer function of the bandpassfilter used to extract the CMN error from the raw data is:

S z ,.

H (S) 2(S+WI) (S+W2)

AZ: W = 0.3 rad/sec, W = 10 rad/sec (3-dB points)1 2

EL: W = 0.5 rad/sec, W = 10 rad/sec (3-dB points)1 2

The corresponding digital filter difference equation is: -'

Yn= (4 +2WT +2W2T WW2T 1 2W2T (Xn -Xn_2) + (8 -2WW 2T2) yn-1

* -(4 -2WIT -21; 2 T +WlW 2 T2 ) Y2"".Y,-2

W2

Note that the (S+W2 ) term is the low pass filter term which

may already be built into the receiver output.

A-2 ,"%,

" *.".."

I.-.

. N N. X..*.. . . . . .

* .4. 44 *~ . .,....

5 --,r4"' s.'",.%.. 'V. ,'q- .' "''' " 'Z" ._', .? ;" k '... - 7 " ..€Z', . , . j ' , -..M '.v, ,.1

I" p'IJ! \a,

...:

I,,.!

I..-41

,Th.. . ..I.. . ..- .-..-% .-..' .